Research map / Reinforcement learning / Offline RL
Transitive RL: Value Learning via Divide and Conquer
Seohong Park, Aditya Oberai, Pranav Atreya, Sergey Levine
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use divide-and-conquer (transitive) value update instead of 1-step TD | offline goal-conditioned RL value learning · puzzle-4x5 (overall) | success rate (%): 19 -> 97 | Improved |
| use divide-and-conquer value update instead of Monte Carlo value learning | offline goal-conditioned RL value learning · puzzle-4x6 (overall) | success rate (%): 37 -> 51 | Improved |
| use divide-and-conquer value update instead of MC | offline goal-conditioned RL value learning · humanoidmaze-giant (overall) | success rate (%): 79 -> 79 | No clear effect |
| use expectile kappa=0.7 instead of 0.5 (behavioral regression) (ablation) | TRL · humanoidmaze-giant | success rate | Improved |
| use expectile kappa=0.7 instead of 0.5 (ablation) | TRL · humanoidmaze-medium and scene-play | success rate | No clear effect |
| use in-trajectory (behavioral) subgoals instead of random subgoals (ablation) | TRL · ablation environments | success rate | Improved |
| add distance-based re-weighting (lambda > 0) (ablation) | TRL · ablation environments | success rate | Improved |
| use TRL instead of prior GCRL methods on standard tasks | offline GCRL · OGBench (10 environments, 50 tasks) | average success rate | Improved |
Closest papers by meaning
- A Simple Reward-free Approach to Constrained Reinforcement Learning (ICML 2022 · similarity 0.88)
- Enhancing Value Function Estimation through First-Order State-Action Dynamics in Offline Reinforcement Learning (ICML 2024 · similarity 0.87)
- Physics-informed Value Learner for Offline Goal-Conditioned Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- Option-aware Temporally Abstracted Value for Offline Goal-Conditioned Reinforcement Learning (NeurIPS 2025 · similarity 0.86)
- Conservative Offline Goal-Conditioned Implicit V-Learning (ICML 2025 · similarity 0.86)
- Critic-Guided Decision Transformer for Offline Reinforcement Learning (AAAI 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
